Ontology based E-Learning System for undergraduate

Students using FPN and HMM.

Upasna Ochani#1, Samruddhi Handore#2, Sameera Prabhune#3, Sneha Jagadale#4, Aparna Lavangade#5
#
Indira College Of Engineering And Management, BE Computer Department, Savitribai Phule Pune University
1
ochaniupasna@gmail.com
2
samhh405@gmail.com
3
sameera.prabhune17@gmail.com
4
sneha8rock@gmail.com
5
aparna.lavangade@indiraicem.ac.in

Abstract— E-learning is used for the distance learning which helps the learner to learn anything from anywhere, anytime. Our system
will provides distance learning for students. The idea is to find the pedagogical objective most adapted to the learner’s requirement
because we can find a course with several pedagogical objectives suggested by a teacher. For creating this system we are using KNN,
Mini batch k-means and FPN algorithm for easy access of content for particular subject and also, we are adding the test for each
subject. The goal of e-learning is to provide the user an e-learning platform a pedagogical content according to user’s requirement.
For this purpose we are using FPN and HMM to find the relevant documents for a user request that expresses the user need. This
system can help administrators and lectures to obtain information about user’s activity involved in eLearning. The objective of the
system is to provide students an on-line learning environment that will meets their expectations regarding the use of electronic media
for on-campus programs and to make more effective use of learning resources. Learner’s Profile can be analysed with the help of pre-
test given by the user. The ability of student is analysed and syllabus for the student is designed with the help of pre-test. E-learning
course content and it’s tests are model using Petri Net with Fuzzy Rules (Fuzzy Petri net),HMM (Hidden Markov Model) is used to
find optimal Learning Path Sequence so that results which are obtained are optimal and according to the user profile. The system
provides online compiler for execution of code and better understanding of it. Hence the system will provide students the opportunity
to develop skills enabling them to work and participate in a technology-rich environment.

Keywords—Personalize learning, KNN, Mini batch K means, FPN, Adaptive learning, etc.

I. INTRODUCTION
Nowadays growing number of popularizations in the World Wide Web promotes e-learning via web. During e-learning the
users can easily share, reuse, and organize the knowledge. E-learning is the process of learning anytime and anywhere with
dynamically changing contents of any course using electronic method. In contrast with traditional teaching method of learning
using scheduled classroom session and the course contents are static in nature. Many educational institutions and organizations
are adopting e-learning. Information about e-learning in the web increases day by day. Year by year the number of accessing the
users count is increased in billion, in 2014 it is $52 billion and in 2015 it is double the time increased. Figure 1 shows the
popularity rising chart in e-learning. The keyword-based search method in e-learning system provides irrelevant information
during many search processes. The goal of e-learning is to give the user of an e learning platform a pedagogical content
customized according to user’s requirement. The idea is to find the pedagogical objective most adapted to the learner’s
requirement because we can find a course with several pedagogical objectives suggested by a teacher.

II. PROBLEM DEFINITION

To create an e-learning system based on FPN and HMM algorithms and the concepts of information retrieval system to find the
relevant documents for a user request that expresses the user need.

III. SYSTEM WORKING

The system is developed in such a way that it is divided into three entities student tutor and admin. The system should report the
activities of tutors and student activities The figure 1 shows that learner gives pre-test then its LPS(Learning Path Sequence) is
planned. Student’s profile is developed based on pre-test. Learning repository is formed using Ontology.
E-learning course content and it’s tests are modelled using Petri Net with Fuzzy Rules (Fuzzy Petri net).HMM (Hidden Markov
Model ) is used to find optimal Learning Path Sequence so that results which are obtained are optimal and according to the user
profile. The figure 1 shows user login, time that has been spent, and the amount of data user is accessing in the course. This
system uses php, javascript, xml, css and html in coding. For the storage this system adds a few tables on the sql database. Role
users involved in this system are the administrator, teacher, and student. When the user logs into the application of E-Learning,
it adds new data into the database that contains the values to the time the user logging in. In addition to user login time data, this
system also records users’ activity in the course. Role of tutor/teacher is to design/edit the syllabus and upload to the system.

Fig. 1 System Architecture

Steps involved:
1. The learning objectives will be submitted by the teacher to expert/admin. Admin will develop the assessments
accordingly.
2. Now database will be manipulated accordingly.
3. This database will be analyzed by the tutor.
4. Tutor will design the curriculum considering difficulty parameters.
5. Now the user will view the required pdf and videos.
6. User will undertake the tests.
7. Tests will be moderated by teacher and the teacher will submit the results to admin.
8. Now admin will post the results and that will be viewed by the student.

IV. ALGORITHMS
A. Fuzzy Petri Nets(FPN)
FPN is nothing but the fuzzy Petri Nets, which is used to trace the performance of the students. The FPN works in two phase
one is Place and another is Transaction. The place is nothing but the source content of the subject you learn and the Transaction
is nothing but the content which is learned by the students. With the help of this algorithm we will let know that how much
student learned and how much is remained. The basic working of FPN is shown in the fig.2.

Fig. 2 FPN
In the above figure, as you see the tokens are nothing but the content of the subject and the paces is the whole subject. FPN
cannot support the variant adjustment of the leaning mechanism. So, follow the constantly changes of the learning mechanism
we use HMM. To increase the level of the student we use this HMM algorithm. The level of the performance of student is
mapped with the given criteria to get into next level we use this. By which we can increase the level of the student as per the
performance of the student.

Role of FPN:
Basic steps involved in the modeling of FPN are, first extract information from experts and/or databases, and then fuzzify the
information from crisp set to fuzzy set by defining membership function. Next step is to construct PN by designing rule base,
and inference rules, and in the last, defuzzifiation of the results.

B. Hidden Markov Model(HMM)

Hidden Markov Model (HMM) is a statistical markov model in which the system being modeled is assumed to be a markov
process with unobserved i.e. hidden states. The hidden Markov model can be represented as the simplest dynamic bayesian
network. The HMM is closely related to earlier work on the optimal nonlinear filtering problem by Ruslan Stratonovich who
was the first to describe the forward and backward procedure. In simpler markov model (like a markov chain), the state is
directly visible to the observer, and therefore the state transition probabilities are the only parameters, while in the hidden
Markov model, the state is not directly visible, but the output (in the form of data or "token" in the following), dependent on the
state, is visible.
Each state has a probability distribution over the possible output tokens. Therefore, the sequence of tokens generated by an
HMM gives some information about the sequence of states; this is also known as pattern theory, a topic of grammar induction.
The adjective hidden refers to the state sequence through which the model passes, not to the parameters of the model; the model
is still referred to as a hidden Markov model even if these parameters are known exactly. A hidden Markov model can be
considered a generalization of a mixture model where the hidden variables (or latent variables), which control the mixture
component to be selected for each observation, are related through a Markov process rather than independent of each other.

C. Mini Batched K-means

Mini Batched K-means is used to create the test for the appropriate subject section. With the help of this algorithm we create the
test for each section and the test questions for each student are get shuffled to get the exact performance of the student. In this
we create the batches of the questions which are get shuffled by taking test multiple times.
The fig. 2 shows the Mathematical expression of Mini Batched K-means.

Fig. 3 Mathematical Expression for Mini Batched K-means

V. RESULTS

Fig. 4 Home Page

Fig. 5 Admin Page

 Fig. 6 Admin-Tutor Page

Fig. 7 Languages Home Page

Fig. 8 Language Page

 Fig. 9 Compiler Page

Fig. 10 Student Performance Report

VI. CONCLUSION
The system can help administrators and tutors to obtain information about users’ activity involved in eLearning. The student
progress report is displayed in weekly format, making it easier for activity assessment. Hence a system of E-learning based on a
HMM and FPN , the concepts of information retrieval systems, by searching the pedagogical content that is optimal for a
Learner based on his profile and a set of pedagogical objectives set by a teacher and with what level the learner must learn.

ACKNOWLEDGMENT

We are grateful to our internal guide Prof. Aparna Lavangade (completed ME from SPPU in computer, now working as
assistant professor at Indira college of engineering and management) for her support and guidance throughout the course of our
Project. We are also grateful to our project coordinator Dr. Deepak Dharrao (completed PHD from JJTU in computer
engineering, now working as assistant professor at Indira college of engineering and management) for his support and guidance
throughout the course of our Project.

REFERENCES
[1] Monika Rani, Ranjana Vyas and O. P. Vyas OPAESFH:Ontology-based Personalized Adaptive E-learning system using FPN and HMM Proc. of the
2017 IEEE Region 10 Conference (TENCON), Malaysia, November 5-8, 2017.
[2] Ayush Sharma A Proposed E-learning System Facilitating Recommendation Using Content Tagging and Student Learning Styles IEEE 978-1-5386-
19223/17/2017.
[3] Vishal Pant, ShivanshBhasin , Subhi Jain Self-Learning system for personalized E-learning.
[4] Dr. S. Suguna, V. Sundaravadivelu, B. Gomathi, A novel semantic approach in E-learning Information Retrieval System 2nd IEEE International
Conference on Engineering and Technology (ICETECH), 17th & 18th March 2016.
[5] M. Suryani and Z. A. Hasibuan The Study of Dynamic Delivery Adaptive Learning Content in E-learning Personalization Using Text Mining and
Ontology Approach ISBN: 978-979-1421-19-5 /13/2013 IEEE.
[6] Nidhi Pandey, Ajay Kumar Learning Algorithm for Intelligent Based E-Learning System 978-1-4673-4529-3/12/$31.00 c 2012 IEEE.
[7] Elena Madalina Jianu, Andrei Vasilateanu Designing of E learning system using adaptivity and gamification 978-1-5386-3403-5/17/$31.00 ©2017
IEEE.